jacquesfize / gmatch4py Goto Github PK

View Code? Open in Web Editor NEW

190.0 190.0 41.0 339 KB

A graph matching library for Python

License: MIT License

Python 19.53% Cython 80.47%

gmatch4py's People

Contributors

Stargazers

Watchers

gmatch4py's Issues

Fixing a failing build on Mojave (OSX 10.14)

Attempting to install GMatch4py for Python 3.5 and Python 3.6 on Mojave in a virtual environment fails with clang reporting:

clang: error: invalid deployment target for -stdlib=libc++ (requires OS X 10.7 or later)

This is because the deployment target isn't being recognized by clang. This can be rectified by doing export MACOSX_DEPLOYMENT_TARGET=10.14 before running the build.

Finally, the FutureWarning reported by Cython:

FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release!

can be resolved by changing the extensions line in setup.py to:

extensions = cythonize([makeExtension(name) for name in extNames], compiler_directives={'language_level': '3'})

Reference: https://stackoverflow.com/a/53992016

GMatch4py-build-error.txt

set_attr_graph_used

I'm not sure if I am misunderstanding this, but is it not possible to only pass node attributes or edge attributes, without necessarily having to pass both of them? From base.py I observe:
def set_attr_graph_used(self, *args, **kwargs): # real signature unknown """ Set graph attribute used by the algorithm to compare graphs. Parameters ---------- node_attr_key : str key of the node attribute edge_attr_key: str key of the edge attribute """ pass

Clearer documentation

Considering that people in python often make use networkx to represent their graphs, the output of the compare() function may seem confusing to some people. The output for the networkx version of graph edit distance leads to single number like: 2.0.
For gmatch4py this is a cost matrix like: [[1. 0. ] [0.17535545 1. ]] (normalized). I thought that it may be interesting to add in your documentation an explanation on how to interpret this cost matrix.

HED returns 0

I have a bunch of graphs I know to be unique, and when I calculate GED using GraphEditDistance it returns non-zero values as expected. However when using HED instead of GraphEditDistance, it outputs only 0.

ged = gm.HED(1, 1, 1, 1) 
result = ged.compare(all_graph, None)
print(*result)

Is my code wrong or is there an issue with the library?

What is the order of the costs in GraphEditDistance?

Hi, in the readme it appears that:

ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1

The costs are adding and removing nodes and edges?
but what is the order of the costs?

Installation failed

Hi, it won't install. (Windows 10)
It generates a long error log, but it seems that this is the most informative part:

ERROR: Failed building wheel for GMatch4py
  Running setup.py clean for GMatch4py
Failed to build GMatch4py
DEPRECATION: Could not build wheels for GMatch4py which do not use PEP 517. pip will fall back to legacy 'setup.py install' for these. pip 21.0 will remove support for this functionality. A possible replacement is to fix the wheel build issue reported above. You can find discussion regarding this at https://github.com/pypa/pip/issues/8368.
Installing collected packages: GMatch4py, Cython
    Running setup.py install for GMatch4py ... error
    ERROR: Command errored out with exit status 1:

Import Issues "munkres"

Having trouble importing gmatch4py. Looks like there's a deeper problem importing the munkres package. I've installed it with pip install munkres and it installed fine. Trying to install if from https://github.com/jfrelinger/cython-munkres-wrapper as suggested in the error report prompts an error about not being about to detect the archive format.

What am I doing wrong?

User entry:

C:\Users\user>python
Python 3.7.5 (tags/v3.7.5:5c02a39a0b, Oct 15 2019, 01:31:54) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import gmatch4py

The error:

C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.7_3.7.1520.0_x64__qbz5n2kfra8p0\lib\importlib\_bootstrap.py:219: UserWarning: To obtain optimal results install the Cython 'munkres' module at  https://github.com/jfrelinger/cython-munkres-wrapper
  return f(*args, **kwds)
Traceback (most recent call last):
  File "gmatch4py\ged\abstract_graph_edit_dist.pyx", line 13, in gmatch4py.ged.abstract_graph_edit_dist
ImportError: cannot import name munkres

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\user\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0\LocalCache\local-packages\Python37\site-packages\gmatch4py\__init__.py", line 4, in <module>
    from .ged.graph_edit_dist import *
  File "gmatch4py\ged\graph_edit_dist.pyx", line 1, in init gmatch4py.ged.graph_edit_dist
  File "gmatch4py\ged\abstract_graph_edit_dist.pyx", line 16, in init gmatch4py.ged.abstract_graph_edit_dist
  File "C:\Users\user\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0\LocalCache\local-packages\Python37\site-packages\scipy\optimize\__init__.py", line 390, in <module>
    from .optimize import *
  File "C:\Users\user\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0\LocalCache\local-packages\Python37\site-packages\scipy\optimize\optimize.py", line 37, in <module>
    from .linesearch import (line_search_wolfe1, line_search_wolfe2,
  File "C:\Users\user\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0\LocalCache\local-packages\Python37\site-packages\scipy\optimize\linesearch.py", line 18, in <module>
    from scipy.optimize import minpack2
ImportError: DLL load failed: The specified module could not be found.
>>>

graph-edit example from README: TypeError: 'method' object is not iterable

Hi there. Already running into the next problem:

$ python3
Python 3.7.3 (default, Dec 20 2019, 18:57:59) 
[GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import networkx as nx
>>> import gmatch4py as gm
/usr/lib/python3.7/importlib/_bootstrap.py:219: UserWarning: To obtain optimal results install the Cython 'munkres' module at  https://github.com/jfrelinger/cython-munkres-wrapper
  return f(*args, **kwds)
>>> g1=nx.complete_bipartite_graph(5,4) 
>>> g2=nx.complete_bipartite_graph(6,4)
>>> ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
>>> result=ged.compare([g1,g2],None) 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "gmatch4py/ged/abstract_graph_edit_dist.pyx", line 179, in gmatch4py.ged.abstract_graph_edit_dist.AbstractGraphEditDistance.compare
  File "gmatch4py/ged/abstract_graph_edit_dist.pyx", line 182, in gmatch4py.ged.abstract_graph_edit_dist.AbstractGraphEditDistance.compare
  File "gmatch4py/helpers/general.pyx", line 21, in gmatch4py.helpers.general.parsenx2graph
TypeError: 'method' object is not iterable

I have no idea why this error occurs.

Graph2Vec error

Hello,
I'm trying to use this library for the first time. In particular, I'm trying to embed my graph using Graph2Vec approach. As there's no much documentation, I'm basically testing functions like a monkey, nevertheless I had this error:

g = nx.barabasi_albert_graph(200, 10)
e = gm.embedding.graph2vec.generate_model(g)

---------------------------------------------------------------------------
_RemoteTraceback                          Traceback (most recent call last)
_RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/joblib/externals/loky/process_executor.py", line 431, in _process_worker
    r = call_item()
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/joblib/externals/loky/process_executor.py", line 285, in __call__
    return self.fn(*self.args, **self.kwargs)
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/joblib/_parallel_backends.py", line 595, in __call__
    return self.func(*args, **kwargs)
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/joblib/parallel.py", line 262, in __call__
    return [func(*args, **kwargs)
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/joblib/parallel.py", line 262, in <listcomp>
    return [func(*args, **kwargs)
  File "gmatch4py/embedding/graph2vec.pyx", line 114, in gmatch4py.embedding.graph2vec.feature_extractor
  File "gmatch4py/embedding/graph2vec.pyx", line 89, in gmatch4py.embedding.graph2vec.dataset_reader
  File "/home/cittaro.davide/.conda/envs/schist/lib/python3.8/site-packages/networkx/classes/function.py", line 75, in degree
    return G.degree(nbunch, weight)
AttributeError: 'int' object has no attribute 'degree'
"""

The above exception was the direct cause of the following exception:

AttributeError                            Traceback (most recent call last)
<ipython-input-23-b5451756c7d9> in <module>
----> 1 g1v = gm.embedding.graph2vec.generate_model(g1)

gmatch4py/embedding/graph2vec.pyx in gmatch4py.embedding.graph2vec.generate_model()

~/.conda/envs/schist/lib/python3.8/site-packages/joblib/parallel.py in __call__(self, iterable)
   1059 
   1060             with self._backend.retrieval_context():
-> 1061                 self.retrieve()
   1062             # Make sure that we get a last message telling us we are done
   1063             elapsed_time = time.time() - self._start_time

~/.conda/envs/schist/lib/python3.8/site-packages/joblib/parallel.py in retrieve(self)
    938             try:
    939                 if getattr(self._backend, 'supports_timeout', False):
--> 940                     self._output.extend(job.get(timeout=self.timeout))
    941                 else:
    942                     self._output.extend(job.get())

~/.conda/envs/schist/lib/python3.8/site-packages/joblib/_parallel_backends.py in wrap_future_result(future, timeout)
    540         AsyncResults.get from multiprocessing."""
    541         try:
--> 542             return future.result(timeout=timeout)
    543         except CfTimeoutError as e:
    544             raise TimeoutError from e

~/.conda/envs/schist/lib/python3.8/concurrent/futures/_base.py in result(self, timeout)
    437                 raise CancelledError()
    438             elif self._state == FINISHED:
--> 439                 return self.__get_result()
    440             else:
    441                 raise TimeoutError()

~/.conda/envs/schist/lib/python3.8/concurrent/futures/_base.py in __get_result(self)
    386     def __get_result(self):
    387         if self._exception:
--> 388             raise self._exception
    389         else:
    390             return self._result

AttributeError: 'int' object has no attribute 'degree'

I'm ready to be warned about the wrong usage of the function above.

Deepwalk is a node embedding algorithm.

Deepwalk is a node embedding algorithm (graph embedding ≠ node embedding). As each run of a node embedding can be rotated with an orthogonal rotation matrix doing any sort of comparison of two node embedding is totally meaningless. I am also somewhat surprised to see large blocks of code from my graph2vec repository.

sum up of the matchings between graphs

Hi,
Do you have a function that gives the node matchings between two graphs ?
Sincerely.

Usage documentation?

Hi!

Thanks for sharing this project! It would be really nice if we can get at least some simple examples of how to use it?

I am mostly interested in testing graph distances with edit distance, but fail to produce an working example.

Regards
Bojan

warning: libstdc++ is deprecated; move to libc++ with a minimum deployment target of OS X 10.9 [-Wdeprecated]

Hi, When I executed this command sudo python3 setup.py install, I encountered the following error.

veradeMacBook-Pro:GMatch4py vera$ sudo python3 setup.py install
Password:
gmatch4py/kernels/weisfeiler_lehman.pyx: cannot find cimported module '..base'
gmatch4py/vertex_ranking.pyx: cannot find cimported module '.base'
gmatch4py/vertex_edge_overlap.pyx: cannot find cimported module '.base'
gmatch4py/mcs.pyx: cannot find cimported module '.base'
gmatch4py/bag_of_cliques.pyx: cannot find cimported module '.base'
gmatch4py/bon.pyx: cannot find cimported module '.base'
gmatch4py/jaccard.pyx: cannot find cimported module '.base'
gmatch4py/ged/graph_edit_dist.pyx: cannot find cimported module '.abstract_graph_edit_dist'
gmatch4py/ged/graph_edit_dist.pyx: cannot find cimported module '..base'
gmatch4py/ged/graph_edit_dist.pxd: cannot find cimported module '.abstract_graph_edit_dist'
gmatch4py/ged/bipartite_graph_matching_2.pyx: cannot find cimported module '..base'
gmatch4py/ged/greedy_edit_distance.pyx: cannot find cimported module '.graph_edit_dist'
gmatch4py/ged/hausdorff_edit_distance.pyx: cannot find cimported module '..base'
gmatch4py/ged/abstract_graph_edit_dist.pyx: cannot find cimported module '..base'
gmatch4py/ged/abstract_graph_edit_dist.pxd: cannot find cimported module '..base'
/anaconda3/lib/python3.7/distutils/dist.py:274: UserWarning: Unknown distribution option: 'long_description_content_type'
  warnings.warn(msg)
/anaconda3/lib/python3.7/distutils/dist.py:274: UserWarning: Unknown distribution option: 'setup_requires'
  warnings.warn(msg)
/anaconda3/lib/python3.7/distutils/dist.py:274: UserWarning: Unknown distribution option: 'install_requires'
  warnings.warn(msg)
running install
running build
running build_py
creating build/lib.macosx-10.7-x86_64-3.7
creating build/lib.macosx-10.7-x86_64-3.7/gmatch4py
copying gmatch4py/__init__.py -> build/lib.macosx-10.7-x86_64-3.7/gmatch4py
creating build/lib.macosx-10.7-x86_64-3.7/gmatch4py/helpers
copying gmatch4py/helpers/__init__.py -> build/lib.macosx-10.7-x86_64-3.7/gmatch4py/helpers
running build_ext
building 'gmatch4py.kernels.random_walk_kernel' extension
creating build/temp.macosx-10.7-x86_64-3.7
creating build/temp.macosx-10.7-x86_64-3.7/gmatch4py
creating build/temp.macosx-10.7-x86_64-3.7/gmatch4py/kernels
gcc -Wno-unused-result -Wsign-compare -Wunreachable-code -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -I/anaconda3/include -arch x86_64 -I/anaconda3/include -arch x86_64 -I/anaconda3/lib/python3.7/site-packages/numpy/core/include -I/anaconda3/include/python3.7m -c gmatch4py/kernels/random_walk_kernel.cpp -o build/temp.macosx-10.7-x86_64-3.7/gmatch4py/kernels/random_walk_kernel.o
warning: include path for stdlibc++ headers not found; pass '-std=libc++' on the command line to use the
      libc++ standard library instead [-Wstdlibcxx-not-found]
1 warning generated.
creating build/lib.macosx-10.7-x86_64-3.7/gmatch4py/kernels
g++ -bundle -undefined dynamic_lookup -L/anaconda3/lib -arch x86_64 -L/anaconda3/lib -arch x86_64 -arch x86_64 build/temp.macosx-10.7-x86_64-3.7/gmatch4py/kernels/random_walk_kernel.o -o build/lib.macosx-10.7-x86_64-3.7/gmatch4py/kernels/random_walk_kernel.cpython-37m-darwin.so
clang: warning: libstdc++ is deprecated; move to libc++ with a minimum deployment target of OS X 10.9 [-Wdeprecated]
ld: library not found for -lstdc++
clang: error: linker command failed with exit code 1 (use -v to see invocation)
error: command 'g++' failed with exit status 1

My env is in anaconda python3.6, Mac 10.14(Mojave)

How does the node attr works?

I compare two graphs as follows:

g1=nx.DiGraph()
g1.add_edges_from([(0,3),(1,3),(2,3)])
g1.nodes[0]['type'] = 'Attack'
g1.nodes[1]['type'] = 'Attack'
g1.nodes[2]['type'] = 'Attack'
g1.nodes[3]['type'] = 'Die'

g2=nx.DiGraph()
g2.add_edges_from([(0,3),(1,3),(2,3)])
g2.nodes[0]['type'] = 'Die'
g2.nodes[1]['type'] = 'Die'
g2.nodes[2]['type'] = 'Die'
g2.nodes[3]['type'] = 'Attack'

ged = gm.GraphEditDistance(1,1,1,1)
ged.set_attr_graph_used('type',None)
result = ged.compare([g1,g2],None)

And the comparing result is
[[0,0],
[0,0]]
So it seems that the ged didn't consider the 'type' attribute?

GMatch4py accuracy for graph edit distance

Hi,
I have a simple code to compare the computed graph edit distance with GMatch4py and also the graph_edit_distance function in networkx.
my code is as follows:

g1=nx.complete_graph(3)
g2=nx.complete_graph(4)
ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
result=ged.compare([g1,g2],None)
print("GMatch4py output:\n", result)
print("networkx output:\n", nx.graph_edit_distance(g1,g2))

and the results are:

GMatch4py output:
 [[0. 7.]
 [7. 0.]]
networkx output:
 4.0

Would you please tell me what does each element of the output matrix of GMatch4py mean? Is 7 the edit distance computed by GMatch4py while the exact edit distance is 4? Is the performance and accuracy of GMatchpy low even for very small graphs or it is just an understanding mistake made by me?
Regards

Quick question: Do not quite understand simple output

This simple example shows a different edit distance compared to networkX's GED. Would be great if someone could help me understand the difference. From my understanding of GED, the edit distance should be 1, not 2.

import networkx
import gmatch4py as gm

g1=nx.complete_bipartite_graph(2,3) 
g2=nx.complete_bipartite_graph(2,3)
g2.add_edge(0, 1)
adj=[(v, neighbors) for v, neighbors in g1.adjacency()]
print(adj)
adj=[(v, neighbors) for v, neighbors in g2.adjacency()]
print(adj)
ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
result=ged.compare([g1,g2],None) 
print(result)

[(0, {2: {}, 3: {}, 4: {}}), (1, {2: {}, 3: {}, 4: {}}), (2, {0: {}, 1: {}}), (3, {0: {}, 1: {}}), (4, {0: {}, 1: {}})]
[(0, {2: {}, 3: {}, 4: {}, 1: {}}), (1, {2: {}, 3: {}, 4: {}, 0: {}}), (2, {0: {}, 1: {}}), (3, {0: {}, 1: {}}), (4, {0: {}, 1: {}})]

[[0. 2.]
 [2. 0.]]

ged_nx = nx.graph_edit_distance(g1, g2)
print(ged_nx)

1.0

How to define substitution cost?

Hi.

I have a question about the substitution cost. When we define
ged=gm.GraphEditDistance(1,1,1,1), I think that the parameters are equal to node_del,node_ins,edge_del,edge_ins. Is there a way to input user-defined substitution cost matrix (e.g., a cost of 1 for a mis-match between two edges A and B, and a cost of 0 if two edges are identical)? Or the substitution costs are computed automatically?

Thank you.

Tran.

Installation error

After running python setup.py install :
Terminal prompted with the following error:
gmatch4py/mcs.cpp:27:20: fatal error: Python.h: No such file or directory compilation terminated. error: command 'x86_64-linux-gnu-gcc' failed with exit status 1

What is the maximum edit distance between two graphs?

Hi,
I have a question about the maximum graph edit distance between two graphs.
If Graph A has M nodes, B has N nodes, what is the maximum graph edit distance calculated by GMatch4py between A and B? Is there an upper bound? Whether the maximum graph editing distance depends on M and N?
Finally, which paper presented the calculation algorithm used in GMatch4py?
Looking forward to reply!

No module named 'gmatch4py

Hi,

I am very new to python and Graphs, but I need these similarity measures in a project.
so my question could be a little retarded, sorry.

I tried to install the package as described. (using command prompt)
git clone https://github.com/Jacobe2169/GMatch4py.git
cd GMatch4py
(sudo) pip(3) install .

The first two commands work, but the third does not.

neither (sudo) or (pip3) are recognized .

Fisrt I tried with (sudo): (sudo) pip(3) install .
Then with pip3 (without (sudo)): pip3 install
Then with pip (without (sudo)): pip install

could you help me please?

thanks a lot.

ModuleNotFoundError: No module named 'gmatch4py.ged.graph_edit_dist'

Hi there! Thanks for making this project. I'm following the setup in the README, but something is wrong. I installed using pip3 install ., then I run python3, at the prompt I type

>>> # Gmatch4py use networkx graph 
... import networkx as nx 
>>> # import the GED using the munkres algorithm
... import gmatch4py as gm
Traceback (most recent call last):
  File "<stdin>", line 2, in <module>
  File "/home/hhrutz/Documents/devel/GMatch4py/gmatch4py/__init__.py", line 4, in <module>
    from .ged.graph_edit_dist import *
ModuleNotFoundError: No module named 'gmatch4py.ged.graph_edit_dist'

Issue with install with sudo

Hello,

I did a fresh install on an Ubuntu 18.04.1 virtual machine. If I follow the steps in the readme all is fine to install. However when I run an example, I get this output:

git clone https://github.com/Jacobe2169/GMatch4py.git
cd to GMatch4py
sudo python3 setup.py install
All is well

I build a simple setup.py to test:

from distutils.core import setup
from Cython.Build import cythonize

setup(
ext_modules = cythonize("simpleGmatch4py.pyx")
)

I create a simpleGmatch4py.pyx:

Gmatch4py use networkx graph

import networkx as nx

import the GED using the munkres algorithm

import gmatch4py as gm

g1=nx.complete_bipartite_graph(5,4)
g2=nx.complete_bipartite_graph(6,4)

ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
result=ged.compare([g1,g2],None)
print(result)
print("Done")

I build inplace:
python3 setup.py build_ext --inplace

I then try to run it and get an error:

python3

import simpleGmatch4py
Traceback (most recent call last):
File "", line 1, in
File "simpleGmatch4py.pyx", line 4, in init simpleGmatch4py
import gmatch4py as gm
File "/usr/local/lib/python3.6/dist-packages/gmatch4py/init.py", line 4, in
from .ged.graph_edit_dist import *
File "gmatch4py/ged/graph_edit_dist.pyx", line 1, in init GMatch4py.gmatch4py.ged.graph_edit_dist
ModuleNotFoundError: No module named 'GMatch4py'

This error is similar to Issue #2. In Issue #2 the virtualenv method works perfectly.

Digging further, If I do the additional steps:

pip3 install numpy networkx scikit-learn ipython cython scipy (Note no sudo here)
python3 setup.py install (Note no sudo here)
A. Errors installing in to /usr/local/lib/python3.6/dist-packages
sudo chmod -R 777 /usr/local/lib/python3.6/dist-packages/gmatch4py
python3 setup.py install (Note no sudo here)
Rerun the test above and it works perfectly
[[ 0. 14.]
[10. 0.]]
Done

While I do get it to work, it is not install/working as expected. I hope these details help.

Thank you again for your amazing work! :)

can i use a edge image as a input to Hausdorff ?

Hi can i use a edge image as a input to Hausdorff ？
Thank you !

Some constructive feed

Hi! incredible promises in your Readme.md (speed through Cython, easy API, etc.)
But I found the library a bit unfinished, so here are some (hopefully) constructive feedback:

Import error. After installing via git clone etc. in clean env, as required by readme.md, and just doing import gmatch4py as gm, I still encounter an import issue:

    from .embedding.deepwalk import *
  File "gmatch4py/embedding/deepwalk.pyx", line 29, in init gmatch4py.embedding.deepwalk
ModuleNotFoundError: No module named 'graph'

Autocompletion not working: maybe it's just in my IDE (Pycharm Pro), but I have no autocomplete of functions signatures, or no doctsrings displayed°. Do you have on your side?
2 bis) I know it takes time, but docs are not ultra clear for some functions (for example, how does .set_attr_graph_used works, how does it actually "use" attributes)
aka 2 bis²) : why does .set_attr_graph_used always took 2 arguments, can't we just use edge attr or node attr ? It seems not, but it's not clear to me why.
TypeError: set_attr_graph_used() takes exactly 2 positional arguments (1 given)

° I think it might be because of compiled files .so. From my researchs, it seems docstrings are dropped by defaullt. But here, they give an option to keep docstrings:
Cython.Compiler.Options.docstrings = True

Have a nice day, I hope my feedback will be useful. Again, thanks for open-sourcing your work!
Nicolas MICAUX

ModuleNotFoundError: No module named 'skipgram'

When I try to import gmatch4py, the below error happens:

ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-1-789a2ae4f894> in <module>
----> 1 import gmatch4py as gm

c:\users\XXX\appdata\local\programs\python\python39\lib\site-packages\gmatch4py\__init__.py in <module>
     13 # Graph Embedding import
     14 from .embedding.graph2vec import *
---> 15 from .embedding.deepwalk import *
     16 from .embedding.node2vec import *
     17 # Helpers import

gmatch4py\embedding\deepwalk.pyx in init gmatch4py.embedding.deepwalk()

ModuleNotFoundError: No module named 'skipgram'

`set_attr_graph_used` is not setting `node_attr_key`

Thanks for a great library.

I'd like to calculate GED using node properties aka attributes. Like another issue, this doesn't seem to work in practice: #26

A starting point might be that the set_attr_graph_used doesn't appear to do anything. I'm afraid I don't know how to fix this, since the function does apparently change the attribute: https://github.com/Jacobe2169/GMatch4py/blob/4fc0a822514c65c0d8b12d090b5b89c0af50ef2a/gmatch4py/base.pyx#L68

ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
ged.set_attr_graph_used("atomicNum",'hi')
ged.node_attr_key

output:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
/var/folders/jh/02165y2n7kq2y5ychxtzcjm40000gn/T/ipykernel_37896/2867113632.py in <module>
      1 ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
      2 ged.set_attr_graph_used("atomicNum",'hi')
----> 3 ged.node_attr_key

AttributeError: 'gmatch4py.ged.graph_edit_dist.GraphEditDistance' object has no attribute 'node_attr_key'

How are node attributes being exploited?

The docs mention that the library supports to incorporate node attributes when computing the distance. Without diving deeply into the code, how is this implemented, that is, how are two feature vectors compared to yield a distance?

Fresh Install doesn't work

Hello!

I'd love to try this out, but the installation appears to be broken. Using Ubuntu18.04 I

git clone https://github.com/Jacobe2169/GMatch4py.git
cd to GMatch4py
sudo python3 setup.py install
All is well

I build a simple setup.py to test:

from distutils.core import setup
from Cython.Build import cythonize

setup(
    ext_modules = cythonize("simpleGmatch4py.pyx")
)

I create a simpleGmatch4py.pyx:

# Gmatch4py use networkx graph 
import networkx as nx 
# import the GED using the munkres algorithm
import gmatch4py as gm

g1=nx.complete_bipartite_graph(5,4)
g2=nx.complete_bipartite_graph(6,4)

ged=gm.GraphEditDistance(1,1,1,1) # all edit costs are equal to 1
result=ged.compare([g1,g2],None)
print(result)
print("Hello")

I build inplace:
python3 setup.py build_ext --inplace

I then try to run it and get an error:

python3
>>import simpleGmatch4py
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "simpleGmatch4py.pyx", line 4, in init simpleGmatch4py
    import gmatch4py as gm
  File "/usr/local/lib/python3.6/dist-packages/gmatch4py/__init__.py", line 4, in <module>
    from .ged.graph_edit_dist import *
  File "gmatch4py/ged/graph_edit_dist.pyx", line 1, in init GMatch4py.gmatch4py.ged.graph_edit_dist
ModuleNotFoundError: No module named 'GMatch4py'

Please help :)

Use of GMatch4py

Hi,
Thank you for GMatch4py which seems to works in my computer.
My question is the following:
in my graph, there 3 attributes, 2 on nodes and one a edges.
If an attribute is a float, is it considered as a weight ?
For example, if on the edges I have the distances as attributes, should I transform this attribute to give the more important weights to the shortest distances ?

Importing Cython 'munkres' module

Hi. When I import gmatch4py there is a warning: for optimal results, the Cython 'munkres' module must be installed. But when entering the given url this is totally deprecated and it is impossible to install. How much does this affect the operation of gmatch4py?

A Question of Toponym Geocoding

Hi，I don't know why my emails to you are always unsuccessful, it seems to be the server's reason, so I have to come here to try to contact you, hoping to contact you！
When I reproduced your code of Toponym Geocoding, I found that the data link you gave in the code link can't be opened.For the data needed for the experiment, can you share with me the link that is available now or share the data with me via email?
My email is 360519152@com.

ModuleNotFoundError during Import

Hello, I just installed gmatch4py using pip install git+https://github.com/Jacobe2169/GMatch4py.git, without any errors showing up. Then when I try to import the library, I always get this Error:

ModuleNotFoundError                       Traceback (most recent call last)
Cell In[1], line 13
     11 from grakel import Graph
     12 from grakel.kernels import RandomWalk, ShortestPath, WeisfeilerLehman, GraphletSampling, SubgraphMatching
---> 13 import gmatch4py as gm

File ~\Documents\ece ntua\διπλωματικη\Notebooks-Scripts\thesis_venv\lib\site-packages\gmatch4py\__init__.py:15
     13 # Graph Embedding import
     14 from .embedding.graph2vec import *
---> 15 from .embedding.deepwalk import *
     16 from .embedding.node2vec import *
     17 # Helpers import

File gmatch4py\embedding\deepwalk.pyx:29, in init gmatch4py.embedding.deepwalk()

ModuleNotFoundError: No module named 'graph'

Is there any other dependency/module that I should manually install??

Thanks

Bug with GMatch4py if start with an other example

Hi,
I noticed that if I don't begin with the examples given in your document, GMatch4py fails and gives:
_

TypeError Traceback (most recent call last)
in
----> 1 result=ged.compare([Gx1, Gx2, Gx3],None)
2 print(result)

gmatch4py/ged/abstract_graph_edit_dist.pyx in gmatch4py.ged.abstract_graph_edit_dist.AbstractGraphEditDistance.compare()

gmatch4py/ged/abstract_graph_edit_dist.pyx in gmatch4py.ged.abstract_graph_edit_dist.AbstractGraphEditDistance.compare()

gmatch4py/helpers/general.pyx in gmatch4py.helpers.general.parsenx2graph()

gmatch4py/graph.pyx in gmatch4py.graph.Graph.init()

TypeError: Expected str, got int

If I begin with your example and do the same thing it works!